search

landskape-ai / triplet-attention

Official PyTorch Implementation for "Rotate to Attend: Convolutional Triplet Attention Module." [WACV 2021]
https://openaccess.thecvf.com/content/WACV2021/html/Misra_Rotate_to_Attend_Convolutional_Triplet_Attention_Module_WACV_2021_paper.html
MIT License
406 stars 49 forks source link

Triplet Attention can be applied to EfficientNet in place SE? #22

Closed osivaz61 closed 2 years ago

digantamisra98 commented 2 years ago

@osivaz61 Yes it can be. It should be just a drop in replacement.

osivaz61 commented 2 years ago

@osivaz61 Yes it can be. It should be just a drop in replacement. I am sorry but I could not understand what you say exactly. You tried to apply TA to Effs and finally get drop in performance?

digantamisra98 commented 2 years ago

@osivaz61 Your question was whether TA can be used in Efficient Net in place of SE? And the answer is yes. Regarding whether it gets better or worse performance, we didn't do any experiments on Efficient Net to confirm.